[Greenwich Registrar of Voters] Musca said she had confidence in the machines’ accuracy. There’s no way you can make a mistake. You color in your ovals and the machine reads it,” she said. “It’s as good as scanning their can of peas (at the supermarket). If they trust the price on their can of peas, they should trust this as well.” <ref>
Unfortunately, Grocery Scanning (like ATM banking) is not the same as optical scan voting. Its not just that the computers are different – the whole “system” is different.
If the store programmed the peas incorrectly, then a customer would notice – their receipt and the money they paid would be incorrect – or the store would notice that they were losing money on peas – or perhaps an observant unbiased clerk that did not have blind faith in the scanner would notice. If just one customer or store employee noticed then the problem would be swiftly corrected and likely correct at least until the next change in price for peas was input by another employee.
Why do registrars say that manual counting of ballots is error prone? Because it is done by people who are inherently unreliable. Why do grocery scanners make mistakes? Because they are programmed and updated by people who are inherently unreliable.
How do we detect and correct problems with grocery scanners? We have checks and balances to offset the errors often made by humans – the store should have them while they input prices and upgrade the system – the receipt and money exchange is another check.
How do we detect and correct problems with voting systems – we test as well as we can beforehand – we independently test memory cards (we should) – we do sufficient random audits after elections. We cannot check receipts – we have something better in Connecticut: the ballots filled out by the voters. There is no money transaction we can check, they voter cannot detect an error and see that it is fixed for subsequent voters.
How do we detect and correct problems with human hand counted audits? We count with best practices that reduce problems with counting in the first place: teams of three or four counters; redundant counting of small batches; counting without knowledge of original machine counts; public observation of procedures. If counts don’t match machine results we count again, more carefully. If they still don’t match we do forensic research on the ballots and machines to determine the reason for the discrepancy and then work to determine the cause. (These are all things we should do. Unfortunately, at this time these are not the sort of things that are always done in Connecticut).
Let us not forget that paying a few cents more for a can of peas is hardly the same as losing democracy. We take voting integrity as a hassle at our peril.
The Secretary of the State made public announcements that the University of Connecticut team planned to collect memory cards for a study into the failure rates of the cards. But even so, LHS staff members collected memory cards that failed during the set up and testing phase of the 2007 election. They did this even as the UCONN team was attempting to collect a random sample to test for the number of failing cards and the manner in which they failed. The actions of LHS of replacing failing memory cards during the set up of the election is not outlined in state security protocols and those actions served to remove failing cards from the UCONN study, it ruined portions of the statistical analysis.